ABSTRACT
Mycoplasma pneumoniae is usually susceptible to macrolides, but macrolide-resistant strains have been found frequently in recent years. Mutations in domain V of the 23S rRNA gene of M. pneumoniae interfere with the binding of macrolides to rRNA and mediate macrolide resistance. In this study, we developed a rapid and inexpensive method that combines nested PCR (nPCR), single-strand conformation polymorphisms (SSCPs), and capillary electrophoresis (CE) to detect macrolide-resistant mutants directly from throat swabs. nPCR was used to specifically amplify M. pneumoniae 23S rRNA gene fragments containing mutations, and amplicons were analyzed by CE-SSCP for macrolide resistance mutations, with results confirmed by sequencing. From January to December 2009, 665 throat swabs were collected in Beijing, China, yielding 110 samples that tested positive for M. pneumoniae by nPCR and serological testing. We randomly selected 64 positive throat swabs for CE-SSCP analysis. The A2063G mutation was found in 57 samples, and a coexisting T2611C mutation was identified in 1 sample. An A2063T mutation was identified in 1 sample. The total mutation rate was 91%. All mutant samples identified by nPCR-CE-SSCP were sequenced. The nPCR-CE-SSCP method could identify macrolide-resistant mutants directly from clinical samples. This is the first report of an nPCR-CE-SSCP assay for the detection of dominant mutations that confer macrolide resistance on M. pneumoniae. This approach would allow clinicians to choose appropriate therapy rapidly and could be used as a screening method for genetic mutations related to antibiotic resistance.
Mycoplasma pneumoniae is a common cause of community-acquired respiratory tract infections, especially in children and young adults. Nearly 80% of the population shows evidence of exposure to M. pneumoniae by young adulthood, and epidemics occur every 3 to 7 years (5, 21, 31). For chemotherapy of M. pneumoniae infection in children, erythromycin (ERY), clarithromycin, and azithromycin (AZM) are considered to be first-choice agents. Because of the toxic effect on children, fluoroquinolones and tetracyclines are not recommended.
M. pneumoniae is usually susceptible to macrolides; however, macrolide-resistant strains have been found frequently in recent years (16, 17, 22, 23, 24). The traditional method for assessing the susceptibility of M. pneumoniae to antibiotics depends on determination of the MICs of various macrolides after the isolation of M. pneumoniae strains. This method uses visual indicators of antibiotic effectiveness and is regarded as a reference method. However, it is time-consuming and labor-intensive; thus, it is less valuable in clinical practice than in research. Previous studies have shown that mutations in domain V of the 23S rRNA gene of M. pneumoniae that interfere with the binding of macrolides to rRNA can mediate M. pneumoniae resistance to certain macrolides (2, 8, 9, 19, 29, 32). Lucier et al. found that the A2063G mutation caused ERY resistance in M. pneumoniae strains (16). Okazaki et al. detected the same mutation in clinical samples resistant to ERY, while the A2064G mutation was also detected in resistant strains (22). The C2617A mutation was found to cause resistance to ERY, AZM, and telithromycin (35). Other mutations were also detected in resistant strains (4, 24). The correlation between mutations and macrolide resistance provides the possibility of direct detection of A2063G, A2064G, C2617A, or similar mutations with the goal of chemotherapy selection.
A number of candidate methods have been used to detect mutations directly, including sequencing of PCR products, restriction fragment length polymorphism (RFLP) analysis, or other mutation detection methods, such as real-time PCR and high-resolution melt analysis (25, 33, 34). The single-strand conformation polymorphism (SSCP) technique is based on changes in the conformation of single-stranded DNA (ssDNA) that are induced by a single-base mutation. This can be detected by highly sensitive electrophoresis methods, such as capillary electrophoresis (CE). A strictly controlled SSCP testing procedure, coupled with CE as the separation tool, could reach nearly 100% sensitivity in mutant detection and could be more sensitive than sequencing (1, 6, 7, 10, 11, 13, 14, 26, 27, 28).
In this study, we developed a nested-PCR (nPCR)-CE-SSCP method for rapid detection of antibiotic-resistant mutants of M. pneumoniae. Using this method, nPCR amplifies fragments of the M. pneumoniae 23S rRNA gene directly from the throat swabs of patients. The nPCR products are analyzed by SSCP, and single-stranded DNA is separated by CE. In this study, amplicons were sequenced, and the results were compared to those of CE-SSCP.
MATERIALS AND METHODS
Samples and strains.From January to December 2009, 665 throat swab samples were sent to the Department of Bacteriology at the Capital Institute of Pediatrics from the affiliated children's hospital. All clinical samples were collected from pediatric patients diagnosed as having pneumonia or a respiratory infection according to clinical symptoms.
The following reference strains were used to assess the specificity of the nPCR: M. pneumoniae FH (ATCC 15531), Mycoplasma salivarium (ATCC 23064), Mycoplasma orale (ATCC 23714), Mycoplasma genitalium (ATCC 33530), Ureaplasma parvum (ATCC 27813), Mycoplasma fermentans (ATCC 19989), Mycoplasma hominis (ATCC 23114), Chlamydophila pneumoniae (ATCC 53592), Mycoplasma penetrans (ATCC 55252), Chlamydia trachomatis (ATCC VR-348B), Streptococcus pneumoniae (ATCC 49619), Staphylococcus aureus (ATCC 29213), Staphylococcus epidermidis (ATCC 35984), Escherichia coli (ATCC 11229), Pseudomonas aeruginosa (ATCC 27853), Haemophilus influenzae (ATCC 49247), Moraxella catarrhalis (ATCC 25238), Mycobacterium tuberculosis (ATCC 27294), Candida albicans (ATCC 10231), and strains of influenza viruses A and B (Flu A and Flu B), respiratory syncytial virus (RSV) subgroups A and B, parainfluenza virus 1 to 3 (PIV1, PIV2, and PIV3), adenovirus (AdV), rhinovirus (RV), and coronavirus (CoV) (stored in the Capital Institute of Pediatrics, Beijing, China).
M. pneumoniae infections were detected by an nPCR as described previously (18), and by a serological test with passive particle agglutination (PA; Serodia-Muco II kit; Fujirebio Inc.). Positive results by both methods indicated M. pneumoniae infection. Some of the positive samples were selected randomly for mutation analysis by nPCR-CE-SSCP in domain V of the 23S rRNA gene.
Amplification of fragments containing resistance mutations.DNA was isolated using the TIANcombi DNA Lyse&Amp PCR kit (Tiangen, Hangzhou, China) in accordance with the manufacturer's instructions. Outer primers were designed to amplify one fragment containing all mutations in the 23S rRNA gene (GenBank accession no. X68422.1) according to previous reports. Two sets of inner primers were designed to amplify two short fragments containing different mutation sites (Table 1). All primers were checked with BLAST (http://www.ncbi.nlm.nih.gov/BLAST/ ) for specificity, using all sequences in GenBank.
Primers for amplification of fragments containing resistance mutations
The 23S rRNA fragments were amplified with outer primers 933F and 933R in 50 μl containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 200 μM each deoxynucleoside triphosphate (dNTP), 10 pmol of each primer, 1 U of DNA polymerase (Promega), and 10 μl of the DNA template. The thermal profile was 3 min at 94°C; 35 cycles of 0.5 min at 94°C, 0.5 min at 55°C, and 1 min at 72°C; and 10 min at 72°C. Ten microliters of the primary PCR product was used for secondary PCR, which was prepared and run using the same reagents and conditions, except that primers 303F and 303R, or primers 342F and 342R, were used. Strict procedures were employed to avoid sample contamination. The M. pneumoniae reference strain was used as a PCR-positive control. Other reference strains known to inhabit the human respiratory tract were used to test the specificity of the PCR. The nPCR products were identified on a 2% agarose gel with ethidium bromide (EB).
CE conditions.CE was performed on a Beckman MDQ automatic CE apparatus equipped with a laser-induced florescent (LIF) detector. The coated silica capillary was 40.2 cm long (effective length, 30.0 cm), with an inner diameter of 50 μm (Yongnian, Hebei, China). The running buffer contained 2% (wt/vol) linear polyacrylamide, 0.4% (wt/vol) polyethylene glycol (M w, 20,000) (PEG 20000), and 1× Tris-borate-EDTA (TBE) (pH 8.35). Each day, the capillary was conditioned by pressure rinsing (344.75 kPa) with deionized water for 2 min, 1× TBE (pH 8.35) for 2 min, and running buffer for 10 min, followed by prerun at 4 kV for 5 min and at 12 kV for 10 min with reversed polarity. Between each two runs, the capillary was pressure rinsed (344.75 kPa) with running buffer for 2 min. The sample was injected into the capillary with pressure (3.45 kPa) for 50 s and was separated at 12 kV of reversed polarity for 20 min.
CE-SSCP analysis.The nPCR product, a 342- or 303-bp amplicon in 3 μl, was mixed with 33 μl of deionized formamide and 1.0 μl of 1 M NaOH. The mixture was heat denatured for 5 min at 95°C, chilled on ice for 2 min, and then mixed with 10 μl SYBR green 1 (1:100) and 4 μl 1× TBE (pH 8.35) before CE separation.
The SSCP patterns of the reference strain and clinical samples were superimposed with 32 Karat software (Beckman Coulter, Brea, CA). Migration profiles of clinical samples that were different from those of the reference strain were regarded as indicating mutations.
CE reproducibility test.The reproducibility of CE was tested by analyzing a mixture of 342-bp and 303-bp nPCR products 10 times in 1 day. The electropherograms of each analysis were superimposed in order to compare the migration times of the peaks.
DNA sequencing.All nPCR products of 342 and 303 bp were sequenced (Saibaite Co., Ltd., Beijing, China), and the DNA sequences were compared to the sequence of M. pneumoniae M129 (GenBank accession no. X68422) using BLAST.
RESULTS
Detection of M. pneumoniae infection.Of 665 throat swabs, 110 tested positive by nPCR of 16S rRNA and PA. Most of the samples were obtained during the patient's first visit to the hospital. We chose 64 samples at random for analysis by nPCR-CE-SSCP for 23S rRNA gene mutations. These samples came from 38 boys and 26 girls, 5 months to 15 years old.
Amplification of mutation sites by nPCR.Amplification products targeting the 23S rRNA, including 342-bp and 303-bp fragments, were obtained by nPCR from all 64 M. pneumoniae-positive samples analyzed. All other reference strains were negative (data not shown).
CE-SSCP separation of DNA strands.The double-stranded DNA (dsDNA) nPCR products could be injected directly into capillaries without purification. As shown in Fig. 1, CE clearly separated dsDNA342 and dsDNA303 amplified from the M. pneumoniae reference strain. CE was also able to separate single-stranded DNAs (ssDNA342-1, ssDNA342-2, ssDNA303-1, and ssDNA303-2) from dsDNA342 and dsDNA303 from the M. pneumoniae reference strain (Fig. 2). The mutated ssDNA342 and ssDNA303 from the clinical samples could be separated clearly from the M. pneumoniae reference strain and wild-type strains by CE (Fig. 3).
Superimposed CE electropherograms of nPCR products. Line A, mixture of dsDNA342 (primers 342F and 342R) with the pBR322-HaeIII DNA marker; line B, mixture of dsDNA303 (primers 303F and 303R) with the pBR322-HaeIII DNA marker; peaks 1 to 5, 184-bp, 192-bp, 213-bp, 234-bp, and 267-bp fragments of the pBR322-HaeIII DNA marker.
Superimposed CE electropherograms of ssDNA from nPCR products. Line A, ssDNA of 342-bp products. Three peaks were observed, corresponding to one dsDNA and two ssDNA peaks. Line B, ssDNA of 303-bp products. Line C, mixture of ssDNAs of 342-bp and 303-bp products. Four peaks, corresponding to ssDNA303-1, ssDNA303-2, ssDNA342-1, and ssDNA342-2, were observed.
Superimposed CE electropherograms of ssDNA303 (A) and ssDNA342 (B) from mutated and wild-type samples. Shown are electropherograms of ssDNA303 from a mutated clinical specimen (line a), the M. pneumoniae reference strain (line b), and a wild-type isolate (line c) and of ssDNA342 from a mutated clinical specimen (line 1), the M. pneumoniae reference strain (line 2), and a wild-type isolate (line 3).
Of 64 clinical samples, 58 showed ssDNA342 migration profiles different from that of the M. pneumoniae reference strain, and 1 also had a different ssDNA303 migration profile. The CE-SSCP patterns of mutated strains could be clearly distinguished from that of the reference strain.
To test the reproducibility of the CE, a mixture of dsDNA342 and dsDNA303 from the reference strain was analyzed 10 times by CE-SSCP. Reproducible results were obtained (Fig. 4).
CE reproducibility test. Shown are superimposed CE electropherograms of a mixture of dsDNA342 and dsDNA303.
Sequencing results for PCR products.The 342-bp and 303-bp products from the 64 samples were sequenced. Some sequencing results are shown in Fig. 5. In 57 samples, an A2063G transition was observed in domain V; among these, a concomitant T2611C mutation was found in 1 sample. One sample contained an A2063T transition. All results were consistent with the CE-SSCP results.
Multiple alignment of the 23S rRNA genes from some clinical samples and M. pneumoniae M129 and FH. Partial sequences of the peptidyltransferase (domain V) from positions 2041 to 2091 and 2562 to 2612 are presented. M. pneumoniae numbering is used. The nucleotide sequence of M129 is given according to GenBank accession no. X68422. The reference strain FH was sequenced, and the results are shown. Dots indicate identical nucleotides. The nucleotides at positions 2063 and 2611 are underlined.
DISCUSSION
M. pneumoniae is a common pathogen in respiratory tract infections in children and young adults. Attack rates among closed or semiclosed populations can be quite high, ranging from 25 to 80% in some settings (5, 21, 31). Our data indicated that 17% (110/665) of children 5 months to 15 years old with respiratory infections in our research were infected with M. pneumoniae in 2009.
Mycoplasmas have no cell walls, so ERY or other macrolides are the first choice for antibiotic treatment, especially for young children.
The rate of resistance to ERY is very high. About 17% (13/76) and 6% (12/195) of M. pneumoniae clinical strains isolated from 2000 to 2004 in different studies in Japan were resistant to ERY (17, 19). A recent report from Japan showed that ERY-resistant strains increased to 31% (37/121) in 2006 (20). In addition, a recent report showed that the ERY resistance rate was 92% (46/50) among M. pneumoniae strains isolated in Beijing, China (34). In a previous study, resistance rates in Shanghai, China, increased from 16.7% in 2005 to 76.5% in 2006 to 100% in 2007 and 2008 (15). In this study, macrolide resistance mutations were found in 91% (58/64) of samples tested. Clearly, macrolide-resistant M. pneumoniae is spreading rapidly in certain parts of Asia.
Macrolide resistance in M. pneumoniae is associated with mutations in 23S rRNA or the ribosomal proteins L4 and L22 (16, 30). A2063G is the most common mutation, followed by A2064G. Other mutations, such as C2617A, and mutations in the ribosomal proteins L4 and L22 are rare (3, 30). Of 13 ERY-resistant M. pneumoniae strains isolated in Japan from 2000 to 2003, 10 (77%) had an A2063G transition, 1 (8%) had an A2064G transition, 1 (8%) had an A2063C transition, and 1 (8%) had a C2617G transition (17). Morozumi et al. reported that of 55 macrolide-resistant strains isolated from 380 M. pneumoniae-positive samples, 50 (90.90%) had an A2063G transition and 5 (9.10%) had an A2064G transition (20). In this study, the predominant mutation was an A-to-G transition (57/64 [89.06%]) at position 2063 in domain V of the 23S rRNA gene, with a coexisting T2611C mutation in one sample (1/64 [1.56%]). One sample contained an A2063T transition (1/64 [1.56%]). The results were similar to those reported by Xin et al. (34).
Previous studies revealed a relationship between mutations and drug resistance in M. pneumoniae, making it possible to determine susceptibility by molecular biology methods, such as sequencing, RFLP analysis, or real-time PCR. However, sequencing and RFLP analysis are time-consuming and expensive and are of no value in clinical practice; they are used only in research. Real-time PCR may be much easier, quicker, and simpler to perform than other methods but may not be able to distinguish between the 2063 and the 2064 mutation without additional simplex real-time PCR, and it cannot detect unknown mutations (12, 17, 25). A rapid and reliable assay for the detection of macrolide resistance in M. pneumoniae would be more clinically valuable than existing procedures. This study reports a method that combines nPCR, CE, and SSCP to detect mutations in M. pneumoniae. CE-SSCP analysis is simple and fast. Compared with the time-consuming method of RFLP analysis or DNA sequencing, the protocol presented here is fast, taking 3 h for nPCR and 30 min for CE-SSCP. Since the CE separation can be finished within 15 min and can be automated, at least 20 samples can be analyzed in 1 day. This method has good sensitivity and reproducibility, and the results of CE-SSCP were consistent with those of DNA sequencing.
nPCR is usually used to increase sensitivity and specificity over those with a one-step PCR technique. In the nPCR method developed here, primers were designed to amplify M. pneumoniae 23S rRNA gene sequences from position 1865 to 2206 (dsDNA342) and from position 2474 to 2776 (dsDNA303), based on related sequences published in GenBank. Primer specificity was verified by BLAST. Negative results for nPCR using other reference strains as templates demonstrated the specificity of the primers. Positive results for nPCR indicated that it can directly detect M. pneumoniae infections from the throat swabs of pediatric patients.
An alternative method for mutational analysis is SSCP, which employs ssDNA conformations as a separation mechanism for the detection of wild-type and mutated DNA. SSCP is a mutation-screening method but normally does not give exact mutation positions by use of the common electrophoresis method. Nonetheless, the A2063G and A2063T ssDNA conformations can be identified using CE because of their specific and unique migration profiles.
CE was crucial for this method. The method applied here, using capillary gel electrophoresis and LIF detection, was newly developed. The separation gel contained linear polyacrylamide and polyethylene glycol 20000 (PEG 20000). The low-viscosity gel could be replaced between runs by a simple pressure-driven rinse, and PEG 20000 decreases electroosmotic flow, shortening migration times and sharpening peaks. CE separates ssDNA with higher resolution than other electrophoresis methods, including separating wild-type and mutated ssDNAs with only a single base pair difference.
In this study, we observed that ssDNA303 had better separation than ssDNA342 from wild-type and mutated strains. SSCP is reported to give better detection sensitivity with shorter DNA segments, as previously reported (28). The CE method used here could differentiate between mutated DNA strands and could be even more powerful if shorter segments were amplified. Nonetheless, PCR product specificity and coverage of the known or possible mutations must always be accommodated. More research is necessary to obtain better resolution with maximal coverage of possible mutations.
The combination of the techniques described above ensured a rapid and sensitive detection method for identifying macrolide-resistant M. pneumoniae. Compared to conventional susceptibility tests or PCR-linked RFLP analysis or DNA sequencing, the nPCR-CE-SSCP is inexpensive and rapid. This assay successfully detected macrolide-resistant M. pneumoniae strains directly in clinical samples and distinguished them from wild-type isolates.
This is the first report of an nPCR-CE-SSCP assay for the detection of mutations that confer macrolide resistance on M. pneumoniae. The approach would allow clinicians to choose appropriate therapies rapidly and could be used as a screening method for genetic mutations related to antibiotic resistance.
As we have noted, our current work has several limitations. Cross-reactivity testing for Mycoplasma lipophilum and Mycoplasma buccale, which are likely to inhabit the human respiratory tract, was not performed. Moreover, questions about the clinical significance of macrolide resistance mutations in M. pneumoniae still exist, since in vitro susceptibility tests for M. pneumoniae do not truly reflect drug effects in vivo. Therefore, further study is necessary to determine the relationship between macrolide resistance mutations of M. pneumoniae and clinical outcomes for patients treated with macrolides.
ACKNOWLEDGMENTS
This work was supported by the fund for excellent talent from the government of Beijing, China.
FOOTNOTES
- Received 9 February 2010.
- Returned for modification 19 April 2010.
- Accepted 10 September 2010.
- Copyright © 2010 American Society for Microbiology
